Magnetically-Assisted Surface Enhanced Raman Spectroscopy (MA

Sep 26, 2014 - CZ.1.07/2.3.00/20.0155) of the Ministry of Education, Youth and Sports of the Czech Republic. This work has also been supported by the ...
19 downloads 0 Views 2MB Size
Article pubs.acs.org/ac

Magnetically-Assisted Surface Enhanced Raman Spectroscopy (MASERS) for Label-Free Determination of Human Immunoglobulin G (IgG) in Blood Using Fe3O4@Ag Nanocomposite Anna Balzerova, Ariana Fargasova, Zdenka Markova, Vaclav Ranc,* and Radek Zboril* Regional Centre of Advanced Technologies and Materials, Department of Physical Chemistry, Faculty of Science, Palacký University in Olomouc, 17. listopadu 12, CZ-77146, Olomouc, Czech Republic S Supporting Information *

ABSTRACT: Development of methods allowing determination of even ultralow levels of immunoglobulins in various clinical samples including whole human blood and plasma is a particular scientific challenge, especially due to many essential discoveries in the fields of immunology and medicine in the past few decades. The determination of IgG is usually performed using an enzymatic approach, followed by colorimetric or fluorimetric detection. However, limitations of these methods relate to their complicated setup and stringent requirements concerning the sample purity. Here, we present a novel approach based on magnetically assisted surface enhanced Raman spectroscopy (MA/SERS), which utilizes a Fe3O4@Ag@streptavidin@anti-IgG nanocomposite with strong magnetic properties and an efficient SERS enhancement factor conferred by the Fe3O4 particles and silver nanoparticles, respectively. Such a nanocomposite offers the possibility of separating a target efficiently from a complex matrix by simple application of an external magnetic force, followed by direct determination using SERS. High selectivity was achieved by the presence of anti-IgG on the surface of silver nanoparticles coupled with their further inactivation by ethylamine. Compared to many recently developed sandwich methods, application of single nanocomposites showed many advantages, including simplicity of use, direct control of the analytic process, and elimination of errors caused by possible nonspecific interactions. Moreover, incorporation of advanced spectral processing methods led to a considerable decrease in the relative error of determination to below 5%. of fluorophore defined by a design of the method can be advantageous for the above-mentioned reasons. However, such approaches may be limited by the modification steps required and by possible photobleaching or autofluorescence. Other well-established methods are based on separation techniques (electrophoresis, liquid chromatography), which are often coupled with mass spectrometry.6 These techniques possess several advantages, including high sensitivity and robustness. However, they also have some drawbacks regarding the requirements for sample purity and composition. An alternative technique that can be used for the determination of human IgG is surface enhanced Raman spectroscopy (SERS),7 which allows detection of various organic, inorganic, and biological compounds at ultralow concentrations, even at the molecular level. Until now, SERS has been successfully applied in many fields, ranging from the enterprise sector, clinical praxis to environmental control, and forensic sciences.8 Moreover, the application potential of SERS can be further increased by employing magnetic nanomaterials

I

mmunoglobulins (Ig) play an essential role in many defensive mechanisms of organisms against various potentially damaging objects, such as viruses or bacteria. Five classes have been shown to be present in the human circulatory system, namely immunoglobulins A, G, M, E, and D.1 It is noteworthy that IgG has a considerably smaller effective diameter compared to other immunoglobulins in the abovementioned classes.2 It is thus able to enter placenta and defend an unborn organism in its prenatal phase of development. IgG is present in the blood of healthy human adults at concentrations of approximately 10 g·L−1, but these levels are usually changed during pathological processes induced by diseases.3 Therefore, advanced technologies and novel materials have been utilized to improve currently used methods to develop new procedures for determining levels of IgG with superior selectivity, even at ultralow concentrations, in order to decrease sample requirements, particularly the amount. Classical methods and approaches for measuring IgG concentrations include immunomethods based on spectroscopic detection, such as enzyme-linked immunosorbent assay (ELISA).4 Several alternatives to ELISA have been developed in order to increase the method accuracy and improve the limit of detection, e.g., fluorescence-based methods.5 The presence © 2014 American Chemical Society

Received: May 20, 2014 Accepted: September 26, 2014 Published: September 26, 2014 11107

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114

Analytical Chemistry

Article

for efficient sample separation and preconcentration,9 an approach referred here as magnetically assisted surface enhanced Raman spectroscopy (MA-SERS). This innovative approach has already been successfully applied for the purification of recombinant biotinylated human sarco-/ endoplasmic reticulum Ca2+‑ATPase using a magnetic nanomaterial functionalized with avidin.10 Another important study showed that a magnetic nanocomposite containing silver nanoparticles could be utilized for the detection of a protein−small molecule complex (avidin−biotin) using SERS.11 MA-SERS has also recently been applied for the determination of dopamine in an artificial cerebrospinal fluid and mouse striatum.12 Another possibility is to use various magnetic nanomaterials with gold nanoparticles.13 Drake et al. exploited the magnetic properties of iron oxide nanoparticles to trap and isolate Staphylococcus aureus bacteria and then employed active gold nanoparticles (AuNPs) coated with 4mercaptobenzoic acid for the detection and quantification of the bacteria by SERS.14 One of the key features of the nanocomposites presented here is the formation of a covalent bond between the magnetic and noble metal nanoparticles. Such a property is useful not only for the analysis of ex-vivo samples because of the increased stability of the nanocomposites but also for future applications in in vivo experiments, where a stable bond between the magnetic core and surface modified silver/gold nanoparticles is crucial. To the best of our knowledge, the detection of proteins by SERS-based immunoassay has so far mostly been indirect and required a Raman label to provide a strong Raman signal.15 An example of such an approach can be seen in the work of Chen et al.16 They developed an analytical approach for detecting IgG at a very low limit of detection (LOD) of 0.1 μg.mL−1.16 Moreover, Song et al. prepared gold nanoparticles labeled with 4-mercaptobenzoic acid and used them as a Raman label for the detection of IgG at a level of 100 fm.mL−1.17 It is worth mentioning that this approach can be enhanced by employing sandwich systems.18 The magnetic properties of some nanoparticles can be employed to improve the separation of targets from a complex matrix. A common sandwich-type analytical system thus consists of a combination of magnetic and SERSactive substrates. The magnetic substrates are used to efficiently extract a target from a complex matrix by employing a selective bond between an analyte and immunorecognition molecule (previously immobilized on the surface) and applying an external magnetic force. The SERS-active silver or gold nanoparticles are added after purification and selectively attached to the target using the same set of immunorecognition molecules present on the metal surface. Indirect analysis is then performed by measuring the signal of a Raman label present as a linker between the antibody and metal surface of the SERSactive substrate.17,19 Chon et al. demonstrated the application potential of such an approach for the analysis of IgG.19b They developed a method based on SERS and opto-fluidics for the detection. The LOD was between 1 and 10 ng·mL−1.19b The above-mentioned methods have many advantages, such as ultralow LODs and versatility. However, their simplicity of use is limited by several drawbacks, including the necessity to synthesize two sets of nanoparticles with limited stability, usually complicated experimental design, high risk of false results as a result of nonspecific interactions between particles and nontargeted compounds attracted from a matrix, and inability to monitor the direct bonding of target to a specific

antibody. These limitations may explain the relatively low uptake of such methods in other scientific fields. Here we present a label-free SERS based method for the determination of human IgG using a single nanocomposite. The method utilized a novel magnetic Fe3O4@Ag@streptavidin@anti-IgG nanocomposite, which comprised a magnetic core modified by O-carboxymethylchitosan (CM) to covalently attach silver nanoparticles. The silver surface was subsequently modified by streptavidin and finally anti-imunoglobulin G (antiIgG). The immobilization of anti-IgG via a bond with streptavidin did not influence its total activity, in contrast to the approaches mentioned above, which utilized unspecific direct immobilization on the metal surface. Application of the method is demonstrated for the determination of human IgG in blood samples obtained by the finger prick method. Despite being relatively straightforward, the proposed method showed high accuracy and reliability. Further, the relative standard errors of determination were reduced to below 5% by using a newly developed advanced spectral processing method, which involved background normalization and utilization of two reference spectral points.



MATERIALS AND METHODS Chemicals. Silver nitrate (p.a.), low molecular weight (LMW) chitosan (75−85% deacetylated), iron(II) chloride tetrahydrate (p.a.) and iron(III) chloride hexahydrate (p.a.), 1[3-(dimethylamino)propyl]-3-ethylcarbodiimide methiodide (EDC), N-hydroxysulfosuccinimide sodium salt (≥98% (HPLC); NHS), IgG from human serum (reagent grade, ≥95% (SDS-PAGE), essentially salt-free, lyophilized powder), antihuman IgG (Fc specific)-biotin antibody produced in rabbit (IgG fraction of antiserum, lyophilized powder), ethylamine (purum, 70% in H2O), streptavidin from Streptomyces avidinii (essentially salt-free, lyophilized powder, ≥13 units/mg protein) and α,ω-bis{2-[(3-carboxy-1-oxopropyl)amino]ethyl}polyethylene glycol (Mr 2000) (carboxy-PEG) were purchased from Sigma-Aldrich and used without further purification. H3PO4 (p.a., 85% w/w) and NaOH (p.a.) were bought from Lach-Ner. Acetic acid (99,8%) and methanol (p.a.) were obtained from P-LAB (Czech Republic). Preparation of standard solutions and buffers. A phosphate-buffered saline solution (PBS; 10 mM, pH 7.5) was prepared from 100 mM stock solution of H3PO4 by fine-tuning its pH value using a highly concentrated solution of NaOH (50%, w/v) under constant stirring. The final concentration was adjusted by addition of a defined volume of water. Stock solutions of 0.2 mol·L−1 NHS and EDC were prepared in a volume of 10 mL weighing exact amounts of each. Solutions were used immediately after preparation. Stock solutions of all used proteins, namely IgG, anti-IgG and streptavidin, were prepared by dissolving 1 mg of protein in 1 mL of water in an Eppendorf tube. Solutions of proteins were kept in the dark at −20 °C for no longer than 5 days. Preparation of Fe3O4@Ag nanocomposite. Magnetic nanocomposites were prepared according to the process described earlier by Markova et al.20 Briefly, magnetic nanoparticles were prepared by the Massart coprecipitation method of an aqueous solution of FeCl3.6H2O and FeCl2.4H2O using sodium hydroxide; 1.5 M NaOH was slowly added to a mixture of FeCl3.6H2O and FeCl2.4H2O to obtain a pH of 11. The synthesis was carried out at room temperature under constant bubbling of nitrogen to prevent further oxidation of the ferrous ions. After 1 h of stirring, the product was washed 11108

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114

Analytical Chemistry

Article

Figure 1. Scheme showing the individual steps in the modification of nanocomposite Fe3O4@Ag.

stirred for 1 h to allow covalent bonds to be formed between between streptavidin and the biotinylated part of the anti-IgG. After washing with PBS buffer, ethylamine at a final concentration of 7% wt. was added to block the rest of the active carboxylic groups. Analysis of real samples. Human whole blood samples (n = 2) were collected by finger prick method performed according to standard operation procedure UOG-REB, SOP011. A sample of volume V = 2 μL was placed on a glass platform, 10 μL of nanocomposite Fe3O4 @Ag@ streptavidin@anti-IgG was immediately added and the mixture carefully stirred. The nanocomposite was magnetically separated from the blood sample and washed twice with deionized water (20 μL). Finally, the nanocomposite was magnetically separated and analyzed using SERS. Apparatus. The ζ potentials of the starting material, intermediate and final products were measured using a Zetasizer NanoZS (Malvern, UK). A transmission electron microscope (TEM) (JEM 2010, Jeol, Japan) was used to obtain images of the unmodified magnetic nanocomposite and final nanocomposites. The TEM was operated at a voltage of 160 kV with a point-to-point resolution of 1.9 Å. Raman spectra were collected using a DXR Raman spectroscope (Thermo Scientific, U.S.A.) equipped with a laser operating at a wavelength of 532 nm. The laser power on the sample was set to 2 mW. Each measured Raman spectrum was an average of 32 experimental microscans. The acquisition time was set to 1s. Infrared spectra were acquired using a Nicolet iS5 infrared spectrometer (Thermo Scientific, U.S.A.). A total of 32 scans were measured and averaged for each material. Raman and IR spectra were evaluated using instrument control software (Omnic, version 8, Thermo Scientific, USA) and highs of target spectral bands were statistically evaluated using LibreOffice (version 4.3.0, The Document foundation, Berlin, Germany).

several times with water and separated using simple magnetic decantation. The magnetic nanoparticles generated were used for the preparation of magnetite-O-carboxymethylchitosan hybrids by adsorption of polymer onto the surface of the asprepared nanoparticles induced by gradually increasing the temperature to 80 °C. Finally, silver nitrate was added to the magnetite-O-carboxymethylchitosan hybrids and silver ions were reduced on the surface of the hybrids by amine groups in the polymer under alkaline conditions and at a temperature of 80 °C. The products were finally washed several times with deionized water and separated using magnetic decantation. Preparation of Fe 3 O 4 @Ag@strepavidin@anti-IgG nanocomposite. Immobilization of anti-IgG on the surface of the previously described nanocomposite Fe3O4@Ag was carried out according to the steps shown in Figure 1. First, 1 mL of magnetic nanoparticles prepared by the process described above was mixed with 100 μL of EDC and NHS (final concentration of both reagents was set to 20 mM) and stirred for 30 min. Next, ethylamine was added to a final concentration of wt.7%. The nanocomposite was collected by application of an external magnetic field and washed twice with 10 mM PBS buffer. Then, 5 mM carboxy-PEG was added (final concentration 1 mM) to the washed magnetic nanocomposite. After 1 h of stirring at 300 rpm, the magnetic nanocomposite was separated from the reaction mixture by simple application of an external magnetic field and washed with PBS buffer. Next, EDC and NHS (1:1, final concentration 20 mM) were added and the reaction stirred for a further 1 h. The resulting nanocomposite was again magnetically separated and washed with PBS buffer. In the next step, nanocomposite with previously activated carboxylic groups was mixed with streptavidin (0.2 mg·L−1) and stirred for 1 h to form covalent bonds between the free amino (−NH2) groups present in the structure of streptavidin and activated carboxylic groups of the nanocomposite. After the addition of streptavidin, the nanocomposite was magnetically separated and washed with 10 mM PBS buffer. Next, nanocomposite with streptavidin was mixed with biotinylated anti-IgG (final concentration 0.2 mg·L−1) and



RESULTS AND DISCUSSION Infrared spectroscopy and ζ- potential measurements were used to monitor changes in the surface characteristics of the as11109

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114

Analytical Chemistry

Article

Figure 2. (A) Measured ζ-potentials of the prepared nanocomposite and of the nanocomposites after each step of modification. (B) IR spectra of the bare nanocomposite and modified nanocomposites. (C) TEM image of Fe3O4@Ag nanocomposite.

stirring for 1 h at 100 rpm. Immobilization of streptavidin onto the nanocomposite surface changed the charge only slightly to −1.7 mV. Successful immobilization of streptavidin was confirmed by measuring IR spectra, which showed an increase of bands interpreted as part of the protein structure (1450 and 1650 cm−1). Streptavidin played a key role for binding the biotinylated part of anti-IgG, which was added to the nanocomposite in the next step (Step 6). The formation of a bond between streptavidin and the biotinylated part of anti-IgG was accompanied by changes in the surface charge to 0.4 mV and by changes in the previously mentioned IR bands. Final blockage of the remaining carboxylic groups with ethylamine was confirmed by an additional increase of the ζ-potential to 0.7 mV (Step 7). Moreover, the ability of the analytical system to bind IgG was tested by IR measurements. It was expected that successful binding of IgG would lead to a further change of protein band intensities, as can be seen in Figure 2C. A TEM image of the as-prepared and surface-modified nanocomposite is shown in Figure 2C, where the bright fringes surrounding the silver nanoparticles (shown in gray) can be interpreted as attached protein layers. Additional TEM images acquired at three important stages of the modification are shown in Figure S1A − Figure S01C. It can be seen that the synthesis did not lead to any considerable structural changes. Moreover, the HRTEM images accompanied by a map with the distribution of Ag and Fe3O4 nanoparticles in the prepared Fe3O4@Ag@ Streptavidin@AntiIgG nanocomposite are included in Figure S1D−S1E. The signal stability of the modified nanocomposite is of the highest importance, and several experiments were conducted in order to evaluate its parameters. The as-prepared nanocomposite Fe3O4@Ag was prepared at day = 0 and stored for a period of 90 days. Its final modifications (labeled as Fe3O4@Ag@Streptavidin@AntiIgG) were prepared at selected instances, namely on days 1, 5, 10, 20, and 90, and their respective Raman spectra were acquired. Average Raman spectra can be seen in Figure S2. The calculated variation of the intensity of the spectral band at 1650 cm−1 between measured spectra is lower than 10%. It can be thus stated that the modification is repeatable and resulting Raman spectra comparable.

prepared nanocomposite during the surface modification process. Such modifications included adsorption of chemicals on the surface of the nanocomposite, activation of functional groups, and formation of new bonds. Infrared spectra and ζpotentials measured during the surface modification process are shown in Figure 2. The measured ζ-potential of an unmodified initial nanocomposite (labeled as Step 0) corresponded to a charge −46.2 mV. The nanocomposite was composed of Fe3O4 and Ag nanoparticles covalently bonded by O-carboxymethylchitosan. This linker was selected because of the presence of free carboxylic groups, which increased the long-term stability of the as-prepared nanocomposite by enhancing its dispersibility in aqueous environments. However, such groups cause relatively high negative surface charge. Therefore, their presence during subsequent surface modification steps had to be minimized to ensure highly selective bonding of streptavidin only to a free surface of silver nanoparticles, while at the same time maximizing the sensitivity of the designed analytical system. The free carboxylic groups were thus removed in the first step by reaction with an aqueous mixture of EDC:NHS, followed by an aqueous solution of ethylamine. The deactivation (Step 1) and blockage (Step 2) of carboxylic groups led to a considerable change of ζ-potential to −20.3 mV. This decrease in negative charge confirmed that the carboxylic groups were successfully activated and efficiently blocked. The residual negative charge was attributed to the bare silver nanoparticles. In the next step (Step 3), free carboxyl groups were generated on the surface of the silver nanoparticles by addition of carboxy-PEG. This step was required for further selective immobilization of streptavidin. The adsorption of carboxy-PEG on the surface of Ag led to a moderate increase of surface charge to −21.1 mV. The difference between the latter value and the initial negative charge (−46.2 mV) was due to different numbers of immobilized functional groups. Activation of the newly created carboxylic groups with an aqueous mixture of EDC and NHS (Step 4) was accompanied by a large decrease in negative charge to a value of −2.1 mV. The following step was performed to immobilize streptavidin on the silver surface (Step 5). Covalent bonds were formed between −NH2 groups of streptavidin and activated carboxyl groups by 11110

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114

Analytical Chemistry

Article

control sample contained several important spectral bands that were attributed to parts of the protein structure of the designed analytical probe. Spectral bands at 1650, 1539, and 1350 cm−1 were interpreted as Amide I, II, and III, respectively. Moreover, the relatively sharp band at 1650 cm−1 indicated an α helix type secondary structure. The band at 605 cm−1 was interpreted as an Ag−N bond vibration originating from the addition of ethylamine in the first steps of the modification of the Fe3O4@ Ag nanocomposite, as can be seen from the experiments summarized in Figure S3. Furthermore, its position and intensity were expected to be insensitive to further formation of a bond between anti-IgG covalently attached to the surface of the analytical probe, and IgG in a sample. On the other hand, successful formation of a bond between anti-IgG and IgG was anticipated to change the absolute intensities and, more importantly, the ratio of the spectral bands of Amide I, II, and III, owing to the different structures of these two proteins. Detailed comparison of spectra of the control and model samples revealed that both the position and intensity of the band at 605 cm−1 remained unchanged, as predicted, whereas the intensities and ratios of the protein spectral bands differed. The selectivity of the here designed method is a very important issue. AntiIgG used in the role of selector has been previously utilized by many authors where a high selectivity of antiIgG toward IgG was successfully demonstrated.16−18 The overall selectivity of this approach was tested using mouse IgG and BSA; the presence of these proteins did not lead to any significant changes in the observed analytical signal, as can be seen in Figure S4; calculated values of RS were not significantly different from the values obtained for the blank solution. In order to evaluate the method’s capabilities for obtaining quantitative information, five calibration samples were prepared. All the calibration samples were measured under the same conditions. The range of concentrations used to construct the calibration curve was selected to comply with demands for methods able to perform ultratrace analysis, and was set from 100 ng·L−1 to 500 mg·L−1. Corresponding Raman spectra of the selected standards are shown in Figure 4.

The as-prepared Fe3O4@Ag@anti-IgG nanocomposite was successively utilized for the determination of levels of one of the most important immunoglobulins, IgG, in human whole blood samples. Based on vibration theory, it was predicted that the specific interaction of IgG with anti-IgG antibody immobilized on the surface of the nanocomposite would lead to a spectral shift of the order of a few wavenumbers or change in intensity of Raman bands corresponding to parts of the binding site of the antibody. To test this hypothesis, standard samples containing 100 ng·L−1 of human IgG dissolved in water were analyzed, as well as blank and control samples. The blank sample contained only the pure unmodified nanocomposite Fe3O4 dispersed in water, whereas the control sample consisted of an aqueous dispersion of nanocomposite Fe3O4@Ag@ streptavidin@anti-IgG without addition of IgG. The spectra obtained are shown in Figure 3. The Raman signal of the

Figure 3. Raman spectra of bare nanocomposite Fe3O4@Ag, modified nanocomposite obtained by immobilization of streptavidin and antiIgG (Fe3O4@Ag@streptavidin@anti-IgG), and model sample with a concentration of IgG 16 ng·L−1 (labeled as Fe3O4@Ag@streptavidin@ anti-IgG@IgG).

Figure 4. Raman spectra obtained in the presence of model samples containing 16 ng·L−1, 160 ng·L−1, and 16 μg·L−1 IgG. 11111

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114

Analytical Chemistry

Article

Figure 5. (A) Raman spectrum obtained in the presence of a real sample of whole human blood. For comparison, the Raman spectra of a model sample containing 16 μg·L−1 IgG and bare nanocomposite are also presented. (B) Concentration determination of analyzed samples. Blue points stand for real samples and brown points show data obtained by standard additions. (C) Calibration curve for a determination of IgG in model samples.

curve, S/N = 3). The limit of detection of most current assays for the analysis of IgG in human blood samples is of the order of tens of μg·L −1 , with a relative standard error of determination of around 20%; this includes serum and whole blood samples. Although those assays are based on similar principles (interaction of the target with a particular antibody), the developed method is capable of detecting human IgG at 1000× lower concentration level and at the same time uses much lower sample volumes. This is due to a superior signal amplification effect achieved by careful selection of the SERS conditions (described in the Materials and Methods). Moreover, 10 measurements were taken on 10 separate days to test the stability of the Raman signal. It was found out that the relative error of RS between replicates was lower than 8%. Next, measurements of two model samples were performed to test the method reliability and robustness. Sample A contained 10 ng·L−1 IgG and Sample B 100 ng·L−1, with differences between the real and experimentally obtained values of around 20% and relative standard deviation, RSD = 5%. The same samples were measured at two different times of day (intraday values) and on three separate days (interday values). It was shown that the relative error was lower than 5% for the intraday experiments and lower than 15% for the interday experiments. This demonstrates the applicability of the described method for the analysis of IgG. The application potential of the proposed method was further tested with the aim of using it for the analysis of human whole blood samples obtained using the finger prick method (details are given in the experimental part), where sample amount is strictly limited to a few microliters. It should be noted that a design of the described procedure was tuned to minimize nonspecific interactions of the synthesized nanocomposites with nontargeted compounds present in the complex matrix of human blood. However, part of the overall Raman signal may still have originated from nonspecific interactions, and thus the Raman signal had to be properly verified and normalized by careful comparison of the spectra obtained for real and model samples. The experimental analysis was performed according to the procedures previously used in the calibration experiments. Briefly, 2 μL of 100× diluted whole

Normalized and baseline-corrected spectra were evaluated and used for a calibration. Presented fluctuations in the spectral data, observed from the instability of the spectral band at 605 cm−1, do not allow a direct determination of the amount of IgG from absolute intensities, as can be seen in Figure 4. However, previous works suggest that protein interactions lead to considerable changes in ratios of selected spectral bands.8 Quantification of IgG is thus based on a simple equation (eq 1) that allows calculation of the ratio of the intensities of the protein 1539 and 1650 cm−1 bands, where the absolute intensities of I1650 and I1539 were previously normalized to the intensity of the reference band at 605 cm−1 in order to minimize the unwanted fluctuations in the analytical signal. R=

I1539/I605 I1650/I605

(1)

where I1539, I1650, and I605 stand for the intensities of the spectral bands at 1539, 1650, and 605 cm−1, respectively. I605 was used as a spectral reference to normalize the whole spectrum in order to minimize system errors originating from changes in the composite constitution, structure, and Raman signal instability. As the data come from separate spectral analyses, values of I605 in eq 1 could not be completely eliminated. Moreover, to make the method more robust, a Raman spectrum for a control sample containing only Fe3O4@Ag@streptavidin@anti-IgG composite was used as a reference, as shown in eq 2: RS =

IStd1539/IStd605

IRef1539/IRef605

IStd1650/IStd605

IRef1650/IRef605

(2)

where IS denotes the intensities of the spectral bands obtained by analysis of the samples and, analogously, IRef stands for the spectral intensities of bands in the Raman spectra of the control samples. RS thus describes the degree of change of the protein band ratios between the control and each analyzed sample; the greater the value above one, the higher the amount of IgG present in the sample (control sample; RS = 1). It was found that the ratio showed a linear trend over the whole measured range with a coefficient of determination amounting to 0.98, as can be seen in Figure 5. The calculated limit of detection was 0.6 ng·L−1 (calculated from the parameters of the calibrations 11112

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114

Analytical Chemistry

Article

for methods able to determine IgG at ultralow concentrations. Here, we report a novel method for the determination of human IgG in whole blood samples. The method utilizes a magnetic Fe3O4@Ag nanocomposite, where silver nanoparticles are covalently bonded using carboxymethylchitosan chosen as a physiologically compatible linker. The surface of the asprepared nanocomposite was modified using streptavidin, followed by anti-IgG. The nanocomposite enabled the development of a robust method for the direct determination of IgG at concentrations from 600 fg.mL−1. The benefit of such an approach is that it can be modified relatively easily in order to analyze many other targets by selecting an appropriate antibody modified by addition of biotin.

blood was collected and deposited on a glass platform. Then, 10 μL of a solution containing prepared Fe3O4@Ag@ streptavidin@anti-IgG nanocomposite was added and well mixed with the sample. The nanocomposite was washed twice with water using an external magnetic force for separation (in total 10 μL) and analyzed using SERS. To perform a full statistical evaluation, four samples from two patients were collected and analyzed 5 times. A SERS spectrum for the real sample is presented in Figure 5A, showing that all the bands previously attributed to the analytical probe were present; nevertheless, the spectral band interpreted as Amid II is moderately shifted due to a different structure of IgG, compared to antiIgG and Streptavidin. The Raman spectrum of the real sample was thus compared to a respective Raman spectrum of a standard solution of human IgG (initial concentration = 5 g·L−1) in order to demonstrate the origin of the analytical signal. Results are shown in Figure S4, where the same pattern of both spectra is demonstrated. Further demonstration of the method involved evaluation of the method selectivity. It can be stated that the immunochemical reaction can be considered as very selective, and the impact of such a statement has already been demonstrated. Moreover, the approach used in this analytical method further minimized possible interferences caused by the presence of nontargeted proteins by the applied synergy of magnetic separation and preconcentration with a high selectivity of AntiIgG. Albeit the number of already published papers dealing with the applications of antibodies in the role of selectors was mentioned previously,15−18 the selectivity of the here described method was further tested on a set of proteins, namely mouse IgG, human IgG, and bovine serum albumin. The results obtained from the performed experiment are shown in Figure S5, where it can be seen that the presence of BSA or mouse IgG did not lead to any considerable changes in the spectra; the calculated value of RS is equal to 1 for the blank solution (without human IgG), 1.05 for BSA (c = 100 mg·L−1), and 1.08 for mouse IgG. Nonetheless, the concentration level of human IgG in real samples was quantified by a standard addition method in order to minimize possible spectral interferences originating from the matrix effects and to achieve a suitable level of accuracy and precision. A defined amount of IgG was thus added to each sample, and the spectral position and intensity of all bands were evaluated and compared to Raman spectra of real samples. The obtained values are shown in Figure 5B, where real samples are labeled as blue points, and two standard additions are labeled as brown points, respectively. Signal stability was evaluated using a set number of technical replicates (N = 5). Obtained Raman spectra of Sample 1 retrieved from patient A are shown in Figure S6. The obtained relative standard deviation between calculated values of RS is lower than 5%. The results of the analysis show that samples from patient A contained 9 g·L−1 of IgG, and analogously, samples obtained from patient B contained 10 g· L−1 of IgG. It is worth mentioning that the determined amounts of IgG are in good compliance with the levels recorded in a healthy population.21



ASSOCIATED CONTENT

S Supporting Information *

TEM images and Raman spectra. This material is available free of charge via the Internet at http://pubs.acs.org.



AUTHOR INFORMATION

Corresponding Authors

*Phone: +420 585634753. Fax: +420 585634761. E-mail address: [email protected] (V.R.). *Phone: +420 585634337. Fax: +420 585634761. E-mail address: [email protected] (R.Z.). Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS The authors gratefully acknowledge support from the Operational Program Research and Development for Innovations European Social Fund (Project No. CZ.1.05/2.1.00/03.0058), New Technologies UP in Chemistry and Biology (Project No. CZ.1.05/3.1.00/14.0302), and the Research Team of the Regional Centre of Advanced Technologies and Materials with a Focus on Unconventional Experimental Techniques in Materials and Optical Research (RCPTM_FRONT) (Project No. CZ.1.07/2.3.00/20.0155) of the Ministry of Education, Youth and Sports of the Czech Republic. This work has also been supported by the Operational Program Education for CompetitivenessEuropean Social Fund (CZ.1.07/2.3.00/ 20.0056), an internal grant of Palacky University in Olomouc (PrF_2014032), and the Technology Agency of the Czech Republic (Project No. TA03011368). The authors would like to thank Jana Straska and Klara Cepe for providing HRTEM and EDX data.



REFERENCES

(1) Yap, P. L. Clinical Applications of Intravenous Immunoglobulin Therapy; Churchill Livingstone: Edinburgh, New York, 1992; p ix, pp 270. (2) Shakib, F. The Human IgG Subclasses: Molecular Analysis of Structure, Function, and Regulation, 1st ed.; Pergamon Press: Oxford, New York, 1990; p xi, pp 316. (3) Virella, G. Medical Immunology, 6th ed.; Informa Healthcare: New York, 2007; p ix, pp 465. (4) Dobson, R.; Topping, J.; Giovannoni, G. J. Med. Virol. 2013, 85, 128−131. (5) van Eeden, P. E.; Wiese, M. D.; Aulfrey, S.; Hales, B. J.; Stone, S. F.; Brown, S. G. A. PLoS One 2011, 6, e16741. (6) de Costa, D.; Broodman, I.; VanDuijn, M. M.; Stingl, C.; Dekker, L. J. M.; Burgers, P. C.; Hoogsteden, H. C.; Smitt, P. A. E. S.; van Klaveren, R. J.; Luider, T. M. J. Proteome Res. 2010, 9, 2937−2945.



CONCLUSIONS Current methods for the determination of levels of IgG in humans are mostly based on interactions of antibodies with a target coupled with colorimetric or fluorimetric detection. These methods are usually robust and fast. However, their further development is limited and there is increasing demand 11113

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114

Analytical Chemistry

Article

(7) Rodriguez-Lorenzo, L.; Fabris, L.; Alvarez-Puebla, R. A. Anal. Chim. Acta 2012, 745, 10−23. (8) (a) McNay, G.; Eustace, D.; Smith, W. E.; Faulds, K.; Graham, D. Appl. Spectrosc. 2011, 65, 825−837. (b) Mikhonin, A.; Ahmed, Z.; Ianoul, A. J. Phys. Chem. B 2004, 108, 19020−19028. (c) Sjöberg, B.; Foley, S.; Cardey, B.; Enescu, M. Spectrochim. Acta, A: Mol. Biomol. Spectrosc. 2014, 128, 300−311. (d) Guerrini, L.; Pazos, E.; Penas, C.; Vázquez, M. E.; Mascareñas, J. L.; Alvarez-Puebla, R. A. J. Am. Chem. Soc. 2013, 135, 10314−10317. (e) Avci, E.; Culha, M. Appl. Spectrosc. 2014, 68, 890−899. (9) Trang, N. T. T.; Thuy, T. T.; Higashimine, K.; Mott, D. M.; Maenosono, S. Plasmonics 2013, 8, 1177−1184. (10) Magro, M.; Faralli, A.; Baratella, D.; Bertipaglia, I.; Giannetti, S.; Salviulo, G.; Zboril, R.; Vianello, F. Langmuir 2012, 28, 15392−15401. (11) Chen, L.; Hong, W.; Guo, Z.; Sa, Y.; Wang, X.; Jung, Y. M.; Zhao, B. J. Colloid Interface Sci. 2012, 368, 282−286. (12) Ranc, V.; Markova, Z.; Hajduch, M.; Prucek, R.; Kvitek, L.; Kaslik, J.; Safarova, K.; Zboril, R. Anal. Chem. 2014, 86, 2939−2946. (13) (a) Quaresma, P.; Osorio, I.; Doria, G.; Carvalho, P. A.; Pereira, A.; Langer, J.; Araujo, J. P.; Pastoriza-Santos, I.; Liz-Marzan, L. M.; Franco, R.; Baptista, P. V.; Pereira, E. RSC Adv. 2014, 4, 3659−3667. (b) Neng, J.; Harpster, M. H.; Wilson, W. C.; Johnson, P. A. Biosens. Bioelectron. 2013, 41, 316−321. (14) Drake, P.; Jiang, P. S.; Chang, H. W.; Su, S. C.; Tanha, J.; Tay, L. L.; Chen, P. L.; Lin, Y. J. Anal. Methods 2013, 5, 4152−4158. (15) (a) Lin, C.-C.; Yang, Y.-M.; Chen, Y.-F.; Yang, T.-S.; Chang, H.C. Biosens. Bioelectron. 2008, 24, 178−183. (b) Chen, J.; Luo, Y.; Liang, Y.; Jiang, J.; Shen, G.; Yu, R. Anal. Sci. 2009, 25, 347−352. (c) Li, T.; Guo, L.; Wang, Z. Anal. Sci. 2008, 24, 907−910. (d) Penn, M. A.; Drake, D. M.; Driskell, J. D. Anal. Chem. 2013, 85, 8609−8617. (e) Chen, J. W.; Lei, Y.; Liu, X. J.; Jiang, J. H.; Shen, G. L.; Yu, R. Q. Anal. Bioanal. Chem. 2008, 392, 187−193. (f) Han, X. X.; Kitahama, Y.; Tanaka, Y.; Guo, J.; Xu, W. Q.; Zhao, B.; Ozaki, Y. Anal. Chem. 2008, 80, 6567−6572. (g) Li, T.; Guo, L. P.; Wang, Z. X. Biosens. Bioelectron. 2008, 23, 1125−1130. (h) Wei, L.; Jin, B.; Dai, S. J. Phys. Chem. C 2012, 116, 17174−17181. (i) Zengin, A.; Tamer, U.; Caykara, T. Biomacromolecules 2013, 14, 3001−3009. (16) Chen, J. W.; Lei, Y.; Liu, X. J.; Jiang, J. H.; Shen, G. L.; Yu, R. Q. Anal. Bioanal. Chem. 2008, 392, 187−193. (17) Song, C.; Wang, Z.; Zhang, R.; Yang, J.; Tan, X.; Cui, Y. Biosens. Bioelectron. 2009, 25, 826−831. (18) Narayanan, R.; Lipert, R. J.; Porter, M. D. Anal. Chem. 2008, 80, 2265−2271. (19) (a) Chen, Y.; Cheng, H.; Tram, K.; Zhang, S.; Zhao, Y.; Han, L.; Chen, Z.; Huan, S. Analyst 2013, 138, 2624−2631. (b) Chon, H.; Lim, C.; Ha, S.-M.; Ahn, Y.; Lee, E. K.; Chang, S.-I.; Seong, G. H.; Choo, J. Anal. Chem. 2010, 82, 5290−5295. (c) Zong, S.; Wang, Z.; Zhang, R.; Wang, C.; Xu, S.; Cui, Y. Biosens. Bioelectron. 2013, 41, 745−751. (d) Yoon, J.; Choi, N.; Ko, J.; Kim, K.; Lee, S.; Choo, J. Biosens. Bioelectron. 2013, 47, 62−67. (e) Baniukevic, J.; Hakki Boyaci, I.; Goktug Bozkurt, A.; Tamer, U.; Ramanavicius, A.; Ramanaviciene, A. Biosens. Bioelectron. 2013, 43, 281−288. (f) Guven, B.; Basaran-Akgul, N.; Temur, E.; Tamer, U.; Boyaci, I. H. Analyst 2011, 136, 740−748. (g) Kim, I.; Junejo, I.-R.; Lee, M.; Lee, S.; Lee, E. K.; Chang, S.-I.; Choo, J. J. Mol. Struct. 2012, 1023, 197−203. (h) Shin, M. H.; Hong, W.; Sa, Y.; Chen, L.; Jung, Y.-J.; Wang, X.; Zhao, B.; Jung, Y. M. Vib. Spectrosc. 2014, 72, 44−49. (20) Markova, Z.; Siskova, K.; Filip, J.; Safarova, K.; Prucek, R.; Panacek, A.; Kolar, M.; Zboril, R. Green Chem. 2012, 14, 2550−2558. (21) Reen, D. J.; Murphy, M. B.; Oconnor, A.; Fitzgerald, M. X. Irish J. Med. Sci. 1981, 150, 265−269.

11114

dx.doi.org/10.1021/ac503347h | Anal. Chem. 2014, 86, 11107−11114